I have a log file ('dread-etc.log'). (...) My goal is to be able to run a perl script against this file so that it pulls out only the data that I have defined ("this is a bad error message"). (...) and write out a new log file called only_the_dread.txt.

There are many ways to do this... Here's a command line one-liner for the lazy:

  perl -ne "print if /this is a bad error message/" dread-etc.log >onl
+y_the_dread.txt
[download]

The part (pattern) in between the slashes is a regular expression, see perlre on how to craft them...

Type "perl -h" for a short description of what the options/switches (i.e. -ne) do — the long version is found in perlrun.

Also note that you might want to use single or double quotes, depending on which OS you're on (generally: single quotes on Unix, double quotes on Windows).

___

(Update: quoted parts of the original question, as it has been deleted)

You'd have to modify the code to use the regular expression almut showed you.

If it's a Linux/Unix system you can use the system's tool grep:

grep PATTERN FILE_TO_SEARCH_IN > FILE_WITH_SEARCH_RESULTS
[download]

More info can be found in the manpage of grep.

a simple perl solution could look like this (using some perl magic):

#!/usr/bin/perl
use strict;
use warnings;

my $pattern = shift @ARGV;

while (<>) {
    print if m/$pattern/;
}
[download]

Save that to a file and execute it this way:

perl script.pl PATTERN FILE_TO_SEARCH_IN > FILE_WITH_SEARCH_RESULTS
[download]

update:
formatting changes
perl example added

I wonder, where your original text has gone?

If you need further assistance, you should have left it there, so others could read your description of your problem.

If the given answers aren't sufficient, you should specify where you encounter your problems.